Krasowski et al. BMC Biochemistry 2011, 12:5 http://www.biomedcentral.com/1471-2091/12/5

RESEARCHARTICLE Open Access The evolution of farnesoid X, vitamin D, and pregnane X receptors: insights from the green- spotted pufferfish (Tetraodon nigriviridis) and other non-mammalian species Matthew D Krasowski1,2*,NiAi3, Lee R Hagey4, Erin M Kollitz5, Seth W Kullman5, Erica J Reschly2, Sean Ekins3,6,7

Abstract Background: The farnesoid X receptor (FXR), pregnane X receptor (PXR), and vitamin D receptor (VDR) are three closely related nuclear hormone receptors in the NR1H and 1I subfamilies that share the property of being activated by bile salts. Bile salts vary significantly in structure across vertebrate species, suggesting that receptors binding these molecules may show adaptive evolutionary changes in response. We have previously shown that FXRs from the sea lamprey (Petromyzon marinus) and zebrafish (Danio rerio) are activated by planar bile alcohols found in these two species. In this report, we characterize FXR, PXR, and VDR from the green-spotted pufferfish (Tetraodon nigriviridis), an actinopterygian fish that unlike the zebrafish has a bile salt profile similar to humans. We utilize homology modelling, docking, and pharmacophore studies to understand the structural features of the Tetraodon receptors. Results: Tetraodon FXR has a ligand selectivity profile very similar to human FXR, with strong activation by the synthetic ligand GW4064 and by the primary bile acid chenodeoxycholic acid. Homology modelling and docking studies suggest a ligand-binding pocket architecture more similar to human and rat FXRs than to lamprey or zebrafish FXRs. Tetraodon PXR was activated by a variety of bile acids and steroids, although not by the larger synthetic ligands that activate human PXR such as rifampicin. Homology modelling predicts a larger ligand-binding cavity than zebrafish PXR. We also demonstrate that VDRs from the pufferfish and Japanese medaka were activated by small secondary bile acids such as lithocholic acid, whereas the African clawed frog VDR was not. Conclusions: Our studies provide further evidence of the relationship between both FXR, PXR, and VDR ligand selectivity and cross-species variation in bile salt profiles. Zebrafish and green-spotted pufferfish provide a clear contrast in having markedly different primary bile salt profiles (planar bile alcohols for zebrafish and sterically bent bile acids for the pufferfish) and receptor selectivity that matches these differences in endogenous ligands. Our observations to date present an integrated picture of the co-evolution of bile salt structure and changes in the binding pockets of three nuclear hormone receptors across the species studied.

Background Examples of ligands for NHRs include a range of endo- Nuclear hormone receptors (NHRs) are transcription fac- genous compounds such as bile acids, retinoids, steroid tors that work in concert with co-activators and co- hormones, thyroid hormone, and vitamin D. A few repressors to regulate gene expression [1,2]. Most of NHRs, such as the pair of xenobiotic sensors, pregnane X the NHRs in vertebrates are ligand-activated, although receptor (PXR; NR1I2; also known as steroid and xeno- some NHRs function in a ligand-independent manner. biotic receptor or SXR) and constitutive androstane receptor (CAR; NR1I3), are activated by structurally * Correspondence: [email protected] diverse exogenous ligands. NHRs share a conserved 1 Department of Pathology, University of Iowa Hospitals and Clinics, Iowa domain structure, which includes, from N-terminus to City, IA, 52242, USA Full list of author information is available at the end of the article C-terminus, a modulatory A/B domain, the DNA-binding

© 2011 Krasowski et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Krasowski et al. BMC Biochemistry 2011, 12:5 Page 2 of 19 http://www.biomedcentral.com/1471-2091/12/5

domain (DBD; C domain), the ‘hinge’ Ddomain,the that 5a (planar) bile alcohols are the likely ancestral ligand-binding domain (LBD; E domain), and a variable ligands (the paleomorphic state) while 5b (bent) bile C-terminal F domain that is absent in some NHRs [2,3]. acids are the derived (apomorphic) phenotype [13,23-25]. NHRs show varying degrees of sequence conservation Common 5b-bile acids include chenodeoxycholic acid across vertebrate species in the LBD that likely reflects, (CDCA; 3a,7a-dihydroxy-5b-cholan-24-oic acid) and at least in part, cross-species variation in the physiologi- cholic acid (CA; 3a,7a,12a-trihydroxy-5b-cholan-24-oic cally important ligands. The xenobiotic sensors PXR and acid), the two dominant primary bile acids in humans CAR show extensive cross-species amino acid sequence and other mammals (Figure 1) [13,24]. divergence across vertebrate species [4,5] in addition to PXRsareactivatedbyastructurallydiversearrayof evidence of positive (Darwinian) selection in the LBD endogenous and exogenous molecules that includes bile from phylogenetic analyses [6,7]. Cross-species differ- salts, steroid hormones, prescription medications, herbal ences in xenobiotic ligands are a possible driving force drugs, and endocrine disruptors [26,27]. PXR regulates for changes in the LBD sequence and structure. We and the transcription of enzymes and transporters involved others have also provided data consistent with the in the metabolism and elimination of potentially harmful hypothesis that the structure of the LBD of NHRs in the compounds, including sulfation of toxic bile acids [28]. NR1 H and NR1I subfamilies may have co-evolved with Previous studies have shown substantial cross-species the endogenous ligands in some species in evolution differences in PXR ligand specificity, including in the [4,6-12]. selectivity for bile salts [4,7,8]. Mammalian PXRs are Bile salts are one class of NHR ligands that show sub- activated by a broad range of bile salt structures (both stantial cross-species differences in chemical structure ancestral and evolutionarily derived), while chicken [13]. Bile salts are amphipathic, water-soluble end-meta- (Gallus gallus)andzebrafish(Danio rerio)PXRsare bolites of cholesterol that facilitate intestinal absorption activated by a structurally narrow range of bile salts of lipids, enhance proteolytic cleavage of dietary pro- [4,6,7,9,29]. We and others have proposed that the evo- teins, and exert antimicrobial activity in the small intes- lution of PXRs has been driven by at least two factors: tine [14,15]. Bile salts exhibit significant structural adaptation to evolutionary changes in bile salt (and per- diversity across vertebrate species [13,16-18] and include haps other endogenous molecular) structure and func- bile alcohols (which have a hydroxyl group on the term- tion as a xenobiotic sensor [4,7,9,12]. Compared with inal carbon atom of the side-chain) and bile acids published X-ray crystallographic structures of human (which have a carboxylic acid group on the side-chain) PXR [30-35], a homology model of the zebrafish PXR is (Figure 1) [15]. Primary bile salts are defined as those predicted to have a smaller ligand-binding pocket (LBP) synthesized by the liver, which is accomplished by a than human PXR, with a flat ligand-binding surface complicated biosynthetic pathway that starts with cho- well-suited to binding the planar 5a-bile alcohols that lesterol [19,20]. are the primary bile salts of zebrafish and other cyprini- Secondary bile salts result from modification of pri- form fish [9]. mary bile salts by host bacteria in the intestine [15]. FXR serves as the major transcriptional regulator of Common alterations in bile acids catalyzed by enzymes bile salt synthesis, in part by controlling the expression from anaerobic bacteria in the intestinal tract include of cytochrome P450 (CYP) 7A1, the rate-limiting deconjugation and dehydroxylation, with one possible enzyme in the synthetic pathway [36]. Mammalian FXRs end result being unconjugated and poorly water-soluble are activated best by primary bile acids CDCA and CA bile acids such as lithocholic acid (3a-hydroxy-5b-cho- [37-39]. FXR is typically expressed at high levels in the lan-24-oic acid; LCA), a compound known to be toxic liver, intestine, kidney, and adrenal glands. A second to mammalian species including humans [15,21]. There FXR, termed FXRb (NR1H5), is found in some animal have been few studies of intestinal bile salts in animals species (although it is a pseudogene in the genome of that utilize bile alcohols, but one study of the Asiatic some mammals such as humans and other primates) carp (Cyprinus carpio) showed that the bile alcohol sul- but does not appear to be involved with bile salt binding fates synthesized in the liver of this species did not or regulation [40]. Throughout this manuscript, FXR undergo hydrolysis in the intestinal tract and thus did refers to NR1H4, or what might be termed FXRa in not generate potentially toxic unconjugated secondary species possessing two FXRs. bile alcohols [22]. The African clawed frog (Xenopus laevis) expresses an Variation in the stereochemistry of the junction unusual FXR (also termed FOR, FXR-like orphan recep- between rings A and B of the bile salt steroid nucleus tor) that has a 33 amino acid insert, not found in mam- determines whether bile salts have a flat (planar) or a malian FXRs, in helix 7 of the ligand-binding domain ‘bent’ orientation. Based on surveys of the bile salt com- (LBD) [41]. We have previously demonstrated that sea position in 1153 vertebrate species, we have proposed lamprey (Petromyzon marinus), zebrafish, and African Krasowski et al. BMC Biochemistry 2011, 12:5 Page 3 of 19 http://www.biomedcentral.com/1471-2091/12/5

Figure 1 Representative bile salts and their structures. All bile salts are derived from cholesterol (topmost structure), illustrated with the carbon atoms numbered and the steroid rings labelled A, B, C, and D. Jawless fish, lobe-finned fish, and a limited number of actinopterygian fish use 5a bile alcohols such as 5a-cyprinol-27-sulfate that have an overall planar and extended structure of the steroid rings (see representation of A, B, and C rings on the right side). Most actinopterygian fish (including medaka and Tetraodon nigrivirdis) use 5b bile acids that have an overall bent structure of the steroid rings. One of the two most common primary salts in mammals is chenodeoxycholic acid

(CDCA), the stem C24 bile acid that has the basic 3a,7a-dihydroxylation pattern. Lithocholic acid is one of the smallest naturally occurring bile acids and results from bacterial enzyme-mediated deconjugation and dehydroxylation of primary bile acids. The sodium and calcium salts of lithocholic acid have very low solubility at body temperature. Additionally, lithocholic acid is toxic in humans and other mammals. Krasowski et al. BMC Biochemistry 2011, 12:5 Page 4 of 19 http://www.biomedcentral.com/1471-2091/12/5

clawed frog FXRs are activated by bile alcohols [9]. Sea allowed for comparison to the structures of the corre- lamprey and zebrafish FXRs are selective for planar 5a sponding mammalian receptors, helping to formulate bile alcohols. The Xenopus FXR isoform 1 was activated hypotheses to rationalize the structural changes that by 5a- and 5b-bile alcohols (paralleling the complex bile have occurred during receptor evolution. Additionally, salt profile of this amphibian [23]) but was poorly acti- we build on our previous docking studies with bile acids vated by bile acids [9,41]. and human VDR (hVDR) and expand structure-activity The vitamin D receptor (VDR; NR1I1) is known to series of bile acids at this receptor. Combining informa- mediate the action of 1,25-dihydroxyvitamin D (calci- tion from docking experiments and cross-species triol), a hormone whose ‘classic’ function is to regulate sequence comparisons, we generated site-directed muta- calcium and phosphorus homeostasis. However, VDRs tions and chimeras to better understand the molecular arenowknowntobeinvolvedinawiderangeofphy- determinants of bile salt activation of VDRs. siological processes including immune system modula- tion, skin development, and regulation of the Results metabolism of toxic compounds [42-45]. Two VDR Bile salt profile of Tetraodon nigriviridis genes have been found in the genome of many actinop- We previously reported a survey of the biliary bile salts terygian fish (a likely by-product of whole genome of 213 fish species from 38 different orders [23]. duplication prior to actinopterygian fish radiation) Included in this survey were six species from Tetrao- [46-48], with VDRa and VDRb both shown to be func- dontiformes, an order of actinopterygian fish that tional in the Japanese medaka (Oryzias latipes)[49]. includes Tetraodon nigriviridis. The primary bile salts Mammalian VDRs are activated at low affinity by a nar- from all six species of Tetraodontiformes examined row range of 5b-bile acids, particularly the toxic second- were the common 5b-bile acids CA and CDCA. This is arybileacidLCAanditsderivatives[7,50-52].VDR illustrated in Additional file 1 by high-performance activation in the intestine has been shown to upregulate liquid chromatography (HPLC) and electrospray ioniza- the expression of enzymes (e.g., CYP3A) that can meta- tion-tandem mass spectrometry (ESI/MS/MS) (Addi- bolizeandreducethetoxicityofLCA[52-56].Wepre- tional file 1: figures S1A-S1C). The primary bile salt viouslydeterminedthattheVDRfromthesealamprey pool of Tetraodon is quite similar to that of many mam- (Petromyzon marinus) was insensitive to activation by a mals (including humans) and birds [13,16,17,24,25]. widearrayofbilesalts,including bile acids and bile alcohols with a range of substituents. We proposed the Ligand selectivity of the Tetraodon FXR hypothesis that activation of VDRs by bile acids is a Analysis of the draft genome of Tetraodon nigriviridis ‘derived’ trait, possibly as an adaptation to evolutionary revealed a single putative ortholog to human FXRa changes in bile salt pathways and vertebrate physiology (hFXR) and another single putative ortholog to human that allow for generation of toxic secondary bile acids PXR (hPXR). We did not find any evidence of an ortholog [6,7,9,10]. to FXRb in the Tetraodon genome. The LBDs of tnFXR In this report, we characterize the ligand selectivity of and tnPXR were cloned and functionally expressed as FXR, VDR, and PXR from the green-spotted pufferfish GAL4/LBD chimeras (concentration-response data in (Tetraodon nigriviridis), an actinopterygian fish that Figure 2 and 3). tnFXR was activated most strongly by synthesizes mainly the 5b-bile acids CA and CDCA [23], GW4064 (Figure 2A), a synthetic ligand known to activate thereby having a bile salt profile similar to most mam- mammalian FXRs with high efficacy [60], and activated mals. We also cloned and characterized VDRs from weakly by T-0901317 (Figure 2B), an agonist of mamma- three additional non-mammalian species: medaka, lian liver X receptors (LXRs; NR1H2 and NR1H3), FXRs, African clawed frog, and chicken (Gallus gallus). In and PXRs [34,61,62]. Conjugated and unconjugated 5b terms of primary bile salts, the medaka synthesizes a bile acids also activated tnFXR, including CDCA, tauro- b mixture of 5 C24 and C27 bile acids [57,58]. The Afri- CDCA, 3-oxo-LCA, and LCA (Figure 2C,D; Additional file can clawed frog has a complicated mixture of C26 and 2). The EC50 of bile acids for activation of tnFXR ranged C27 bile alcohols, C24 bile acids, and C27 bile acids in its from 8.8 to 37.6 μM, with maximal effects varying from bile [9,23]. The chicken synthesizes mainly CA and 15-61% of that produced by GW4064 (which served as the CDCA as primary bile salts [25]. reference compound for comparisons of maximal effect or We generated homology models of Tetraodon FXR efficacy). The planar bile alcohol 5a-cyprinol sulfate (the (tnFXR) and PXR (tnPXR) using rat FXR [59] and main bile salt of cypriniform fish but only a very minor human PXR [32] crystallographic structures, respec- component of Tetraodon bile salts) weakly activated tively, as templates. We then performed computational tnFXR. A 76-compound library of known NHR ligands ligand docking experiments to these receptors. This was screened for additional activators of tnFXR. This Krasowski et al. BMC Biochemistry 2011, 12:5 Page 5 of 19 http://www.biomedcentral.com/1471-2091/12/5

Figure 2 FXR concentration-response curves. Ligand activation of FXRs from different species. A) The synthetic mammalian FXR activator GW4064 activates human FXR (hFXR) with high efficacy and submicromolar potency. GW4064 also activates Tetraodon FXR (tnFXR) and zebrafish FXR (zfFXR) but not Xenopus laevis FXR(xlFXR). B) T-0901317 activates hFXR, zfFXR, and tnFXR (although with lower efficacy compared to GW4064) but not xlFXR. C) The primary bile acid chenodeoxycholic acid (CDCA) activates hFXR and tnFXR but not zfFXR or xlFXR. D) The secondary bile acid lithocholic acid (LCA) activates hFXR, tnFXR, and zfFXR, but not xlFXR. The ordinate indicates fold induction compared to vehicle control in luciferase-based assay. Note that the scale of the ordinate is different in A) through D). screening identified three activators of tnFXR in addition bile acids with one or two hydroxyl groups on the to bile acids which included two farnesol derivatives (far- nuclear rings (Figure 3; Additional file 2). Unlike tnFXR, nesol and S-farnesyl-L-cysteine methyl ester) and 6-formy- tnPXR was not activated by GW4064 or farnesol com- lindolo-[3,2-b]-carbazole. pounds. tnPXR was also activated by a variety of andros- tane, estrane, and pregnane steroids, similar to PXRs Ligand selectivity of the Tetraodon PXR from mammals, chicken, and zebrafish (Figure 3C) [4]. With respect to bile salts, tnPXR had similar structure- tnPXR was not activated by several xenobiotics known activity relationships to tnFXR, namely activation by 5b- to be agonists of mammalian PXRs including hyperforin Krasowski et al. BMC Biochemistry 2011, 12:5 Page 6 of 19 http://www.biomedcentral.com/1471-2091/12/5

Figure 3 PXR concentration-response curves. Ligand activation of PXRs from different species. A) T-0901317 activates human PXR (hPXR) and chicken PXR (chPXR) but not Tetraodon PXR (tnPXR) or zebrafish PXR (zfPXR). B) The primary conjugated bile acid taurochenodeoxycholic acid (TCDCA) activates hPXR and tnPXR but not chPXR or zfPXR. C) The steroid 5b-pregnane-3,20-dione (5b-pregnanedione) activates all four PXRs shown. D) Nifedipine activates hPXR, chPXR, and zfPXR, but not tnPXR. The ordinate indicates fold induction compared to vehicle control in luciferase-based assay. Note that the scale of the ordinate is different in A) through D).

(pharmacologically active component of the herbal anti- acids for transactivation of GAL4/LBD constructs of depressant St. John’s wort), nifedipine (Figure 3D), hVDR and mouse VDR (mVDR). LCA and derivatives rifampicin, and SR12813, but was activated by n-butyl (e.g., 3-oxo-LCA, LCA acetate) activated both hVDR 4-amino benzoate, an agonist of Xenopus laevis PXRs and mVDR (Figure 4; Additional file 3). Iso-LCA, with a [4,63] (Additional file 2). single 3b-hydroxy group on the steroid rings (as opposed to 3a-hydroxy of LCA), weakly activated hVDR Structure-activity relationship of bile acids for activation and mVDR. In contrast, naturally occurring 5b-bile of human and mouse VDRs acids with two or three hydroxyl groups on the nuclear Expanding on previous studies from our group [6,7,10] rings were inactive including CDCA (3a,7a-dihydroxy- and others [50,52], we determined the structure-activity 5b-cholan-24-oic acid), DCA (3a,12a-dihydroxy), and relationships of naturally occurring and synthetic bile CA (3a,7a,12a-trihydroxy). We also tested three 5b-bile Krasowski et al. BMC Biochemistry 2011, 12:5 Page 7 of 19 http://www.biomedcentral.com/1471-2091/12/5

Figure 4 Structure-activity for human and mouse VDR. Structure-activity data for bile acid activation of human and mouse VDRs. The bile acids that activate human and mouse VDRs are small, hydrophobic bile acids that are related to lithocholic acid (LCA) and have either a hydroxyl or oxo group at C-3 (LCA, 3-oxo-LCA, LCA acetate, iso-LCA) or have unsubstituted steroid ring (5a-cholanic acid). Addition of a 7a- hydroxy group to LCA (resulting in the bile acid CDCA) or even single steroid ring substitutions at C-7 or C-12 result in bile acids that do not activate the VDRs. Nor-LCA, which has a shortened bile acid side-chain, is also inactive. Krasowski et al. BMC Biochemistry 2011, 12:5 Page 8 of 19 http://www.biomedcentral.com/1471-2091/12/5

acids not known to occur naturally in bile that have a single hydroxyl substituent on the steroid rings on a car- bon other than C-3 (7a-hydroxy, 7b-hydroxy, and 12a-hydroxy-5b-cholan-24-oic acids), as well as unsub- stituted 5b-cholanic acid (no hydroxyl groups on any of the steroid rings). All four of these bile acids were inac- tive with respect to activation of hVDR and mVDR. Thus, bile acids with hydroxyl groups at the C-7 or C-12 position are unfavourable for activation of hVDR (Figure 4). Unsubstituted 5a-cholanic acid, which would have an overall planar orientation of the steroid rings, weakly activated hVDR and mVDR. Two 5a-cholanic acid derivatives (3b-hydroxy and 3-oxo) were inactive (Additional file 3).

Activation of non-mammalian VDRs by bile salts The African clawed frog VDR (Xenopus laevis VDR; xlVDR) was not activated by any bile salts tested, including bile alcohols. In contrast, chicken VDR (chVDR), medaka VDRa (olVDRa), Tetraodon VDRa (tnVDR), and zebrafish VDRa (zfVDRa) were each acti- vated by LCA and/or its derivatives (3-keto-LCA and LCA acetate) but not by bile acids with two or more hydroxyl groups such as CDCA, DCA, or CA (Figure 5 and 6; Additional file 3). The efficacies of LCA, Figure 5 Transactivation of full-length teleost VDRs. HepG2 cells a 3-oxo-LCA, and LCA acetate (in comparison to 1,25 - were transiently transfected with pRL-CMV, XREM-Luc and either dihydroxyvitamin D3) for activation of chicken, medaka, medaka VDRa-pSG5, zebrafish VDRa-pSG5, or Tetraodon VDRa-pSG5 Tetraodon, zebafish VDRs were lower than for hVDR as described in Methods. Cells were exposed to 100 μM of either and mVDR (Figure 6; Additional file 3). lithocholic acid (LCA), 3-keto-LCA, or LCA acetate for 24 hours. VDR response was measured via dual-luciferase assays. Data is represented as the mean fold induction normalized to control Structure-directed mutagenesis experiments (DMSO) ± SEM. We previously used molecular modelling computational docking studies to understand the structural basis of bile acid activation of hVDR and mVDR [9]. These stu- differences in activation of VDR by bile salts. These dies predicted an electrostatic interaction between Arg- mutations were performed in mVDR, which generally 274 (hVDR numbering) and the bile acid side-chain, has higher maximal activation by bile acids but shows a and a hydrogen bond between the 3a-hydroxyl group of similar selectivity for bile acids to hVDR. Three residues, LCA and His-397 in helix 11 (note corresponding resi- previously identified by the hVDR docking model as key due numbers are 5 lower for mVDR; e.g., Arg-269 in to bile acid activation - Arg-269 (R269; charge clamp to mVDR is equivalent to Arg-274 in hVDR). This hydro- carboxylic acid group on bile acid side-chain), His-392 gen bonding brings LCA close to the activation helix 12 (H392;hydrogenbondto3a-hydroxy group of LCA), where LCA forms hydrophobic contacts with Val-398 Phe-417 (F417; stabilization of helix 12) - were mutated and Phe-422 that would stabilize the helix in the opti- individually to two other amino acid resides. This pro- mal orientation for coactivator binding. Site-directed duced a total of six-directed mutants: R269A, R269E, mutagenesis by Adachi et al. supported this conclusion H392A, H392Y, F417A, and F417D. and indicated that alteration on this Arg residue of We also made site-directed mutations at amino acid hVDR (e.g., Arg274Leu) significantly disrupted the positions that showed differences between VDRs that receptor response to LCA [51]. Additional file 4 displays are bile acid-sensitive (human, mouse, chicken, green- the surface around the ligand binding pocket of hVDR, spotted pufferfish) and bile acid-insensitive (African showing that it is predominantly hydrophobic in the clawed frog, sea lamprey, zebrafish). These mVDR muta- middle with more polar features on its ends. tions changed the amino acid residue(s) in mVDR to the We next performed site-directed mutagenesis experi- corresponding residue(s) found in the bile-salt insensi- ments to confirm the docking model of the bile acid to tive VDRs: Q286E (to corresponding residue found in VDR, and to try to rationalize the cross-species xlVDR), S293 D (xlVDR), S293G (sea lamprey VDR), Krasowski et al. BMC Biochemistry 2011, 12:5 Page 9 of 19 http://www.biomedcentral.com/1471-2091/12/5

Figure 6 VDR concentration-response curves. Ligand activation of VDRs from different species, chimeric human-frog VDRs, and mouse VDR

site-directed mutants. A) The plot shows activation of human, chicken, Xenopus laevis, and Tetraodon VDRs by 1,25-dihydroxyvitamin D3. Also shown are activation data for a chimeric receptor that is mostly human VDR (hVDR) with the H1-H3 insert replaced by the corresponding sequence from Xenopus laevis VDR (xlVDR) and the converse chimeric receptor that is mostly xlVDR with the H1-H3 insert replaced by the corresponding sequence from hVDR. B) The plot shows activation of mouse VDR (mVDR) and six site-directed mutants of mVDR (R269A, R269E,

H392A, H392Y, F417A, and F417D) by 1,25-dihydroxyvitamin D3. C) Lithocholic acid (LCA) activates human, chicken, and Tetraodon VDRs, as well as the chimeric receptor hVDR with xlVDR h1-H3 insert, but does not activate xlVDR and the chimeric receptor xlVDR with the hVDR H1-H3 insert. D) Lithocholic acid strongly activates mVDR, weakly activates the site-directed mutants F417A and F417D, and does not activate R269A, R269E, H392A, and H392Y. The ordinate indicates fold induction compared to vehicle control in luciferase-based assay. Note that the scale of the ordinate is different in A) through D).

N319 D (lamprey VDR), N319K (zfVDR), and RCR363- hVDR that lacked amino acid residues 165-218 (hVDR/ 365LCK (xlVDR), RCR363-365RIQ (zfVDR), RCR363- Δins) and an hVDR construct where the insert domain 365ACR (lamprey VDR) (see Additional file 5). Lastly, was replaced with the insertion domain from the bile we also tested the role of the H1-H3 ‘insertion’ domain acid-insensitive xlVDR (hVDR/xlVDRins). The analo- in mediating bile acid activation. Thus, we generated gous constructs were also created for xlVDR: one Krasowski et al. BMC Biochemistry 2011, 12:5 Page 10 of 19 http://www.biomedcentral.com/1471-2091/12/5

Figure 7 Pharmacophore models of Tetraodon VDR, PXR, and FXR. Pharmacophore models were developed for A) Tetraodon VDR (tnVDR), B) Tetraodon PXR (tnPXR), and C) Tetraodon FXR (tnFXR) using the HipHop method in Catalyst™. The structure-activity data used to develop the pharmacophore models are summarized in Additional File 6. A) The tnVDR pharmacophore contains 4 hydrophobes (cyan) and two hydrogen

bond acceptors (green). 1,25-Dihydroxyvitamin D3 (calcitriol) was mapped to the pharmacophore. B) The tnPXR pharmacophore contains three hydrophobes (cyan) and one hydrogen bond acceptor (green). Taurochenodeoxycholic acid was mapped to the pharmacophore. C) The tnFXR pharmacophore contains four hydrophobes (cyan), one hydrogen bond acceptor (green), and one negative ionizable feature (blue). GW4064 was mapped to the pharmacophore. Krasowski et al. BMC Biochemistry 2011, 12:5 Page 11 of 19 http://www.biomedcentral.com/1471-2091/12/5

H3 insertion domain does not appear to play a major role in selectivity for bile acids.

Pharmacophore models for Tetraodon VDR, PXR, and FXR Using the structure-activity data for tnVDR, tnPXR, and tnFXR (summarized in Additional File 6), we developed common features pharmacophore models for these three receptors (Figure 7). The tnVDR pharmacophore, mapped onto 1,25-dihydroxyvitamin D3 in Figure 7A, has four hydrophobes and two hydrogen bond accep- tors. The tnPXR pharmacophore, mapped onto tauro- chenodeoxycholic acid in Figure 7B, shows three hydrophobes and one hydrogen bond acceptor, a smaller pharmacophore than that determined for hPXR [64]. The tnFXR pharmacophore, mapped onto GW4064 in Figure 7C, shows four hydrophobes, one hydrogen bond acceptor, and one negative ionizable feature. Figure 8 PXR-ligand interactions. Interaction map between the secondary bile acid lithocholic acid and PXRs. Key residues involved in binding are listed for human, Tetraodon (tn), and zebrafish (z) Homology modeling of Tetraodon PXR PXRs (the residue number is for human). Conserved resides Structural studies of the LBD of hPXR reveal a large throughout the three species are highlighted by black boxes. The (~1,300 Å3), roughly spherical, hydrophobic LBP with the hydrogen bond is shown by a green dashed line, the electrostatic flexibility to accommodate large molecules such as hyper- charge interaction is presented as a blue dashed line, and the hydrophobic interaction is depicted as an orange dashed line. The forin and rifampicin [30,32-34]. We can also see this by sur- ligand is shown as ball and stick presentation and colored face analysis of the co-crystallized ligands that cover a according to atom types (gray = carbon, red-oxygen). molecular weight range of 273-714 Da and a calculated ALogP (measure of hydrophobicity) range of 3.54-10.11 [65]. The homology model we generated of the LBD of lacking the insertion domain (xlVDR/Δins) and the tnPXR showed an LBP predicted to be slightly smaller other with the insertion domain from hVDR (xlVDR/ (1,230 Å3) compared to hPXR (Figure 8), but larger than hVDRins). the estimated volumes of homology models of the LBPs of The two mutations at Arg269 (R269A and R269E) zebrafish PXR (~1,000 Å3)andXenopus laevis PXRa (~860 reduced the apparent affinity of 1a,25-dihydroxyvitamin Å3)thatwereportedpreviously[9,66].Thehomology D3 by over two orders of magnitude and abolished acti- model results for tnPXR are consistent with the pharmaco- vation by LCA and 3-oxo-LCA. Similar effects were pro- phore model for this receptor described above (Figure 7B). duced by the two mutations at His392 (H392A and We performed molecular docking studies with two ster- H392Y) which reduced the apparent affinity of 1a,25- oids (5b-pregnan-3,20-dione and 5a-androstan-3a-ol) and dihydroxyvitamin D3 by approximately 10-fold and one bile acid (CDCA) that activated tnPXR in the lucifer- 100-fold, respectively, and abolished activation by LCA ase assays (Figure 9). Docking studies show that the two and 3-oxo-LCA. The two mutations at Phe417 (F417A steroids bind similar to the published crystallographic and F417D) both reduced the apparent affinity of 1a,25- structure of 17b-estradiol bound to hPXR [35]. The carbo- dihydroxyvitamin D3 by approximately two orders of nyl group of pregnanedione or hydroxyl group of andros- magnitude while retaining activation by LCA and 3-oxo- tanol at the C-3 position has a hydrogen bond with Thr- LCA but not LCA acetate. Interestingly, unsubstituted 111, which corresponds to Ser-247 in hPXR. For pregna- 5a-cholanic acid activated the R269A, R269E, H392A, nedione, the carbonyl group at C-20 forms another hydro- H392Y, F417A, and F417D mutants weakly. gen bond with Gln-305, as shown in Figure 6. Molecular In general, the mutations at Glu-286, Ser-293, docking indicated that the electrostatic interactions Asn-319, and Arg/Cys/Arg (363-365) had little effect on between CDCA and the conserved Arg-151 residue on activation by 1a,25-dihydroxyvitamin D3 or bile acids helix 5 are important for binding to tnPXR. The hydroxyl compared to wild-type mVDR. Similarly, the four con- groupattheC-3positionformsahydrogenbondwith structs that involved either deletion (hVDR/Δins and Glu-149, which is different from tnFXR, for which a His xlVDR/Δins) or swapping of the H1-H3 insertion residue is involved in hydrogen bonding. For tnPXR, Tyr domain (hVDR/xlVDRins and xlVDR/hVDRins) had lit- is in the position of this His residue, and adopts a different tle effect on activation by 1a,25-dihydroxyvitamin D3 or side-chain rotamer orientation to provide van der Waals bile acids compared to wild-type hVDR. Thus, the H1- contacts with the hydrocarbon scaffold of CDCA. Krasowski et al. BMC Biochemistry 2011, 12:5 Page 12 of 19 http://www.biomedcentral.com/1471-2091/12/5

Figure 9 Homology model of Tetraodon PXR. Plot of interactions of ligands with tnPXR: A) the steroid 5b-pregnan-3,20-dione (pregnanedione) and B) the bile acid CDCA. The legend indicates the types of amino acid residues and amino acid-ligand interactions.

Homology modeling and docking studies of zebrafish van der Waals contacts with A-ring and B-ring of the PXR [9] show an important amino acid sequence differ- hydrocarbon scaffold of bile salts, specifically the methyl ence (Met-243 in human PXR versus Phe at the corre- group at the C-10 position. In zebrafish PXR, the bulky sponding position in zebrafish PXR). This residue is and more rigid benzyl side-chain of the phenylalanine located at the bottom of the PXR LBP and has direct significantly narrow this portion of the zebrafish PXR Krasowski et al. BMC Biochemistry 2011, 12:5 Page 13 of 19 http://www.biomedcentral.com/1471-2091/12/5

LBP compared with human PXR which has the flexible side-chain of methionine-243, which may lead to prefer- ence for planar bile alcohols by zebrafish PXR. tnPXR has valine at the position corresponding to methionine- 243 in human PXR, which allows enough room for either bent or planar formations of the A/B ring of bile salts.

Homology modeling of Tetraodon FXR For the homology model of tnFXR, molecular docking studies show that GW4064 adopts a similar orientation in the structural model as in the hFXR crystal structure [60]. The carboxyl group of GW4064 forms a strong electrostatic interaction with the conserved Arg residue in helix 5 of tnFXR. In addition, extensive hydrophobic contacts occur between the di-chloro-substituted benzyl Figure 10 FXR-ligand interactions. Interaction map between the ring of GW4064 and residues in helices 5, 7, 11 and 12 primary bile acid chenodeoxycholic acid and FXRs. Key residues of tnFXR, which would stabilize the active conformation involved in binding are listed for human, mouse (m), Tetraodon (tn), of receptor. CDCA also activated tnFXR in our experi- and sea lamprey (sl) FXRs (the residue number is for human). ments, but with weaker activity compared with its activ- Conserved resides throughout the four species are highlighted by black boxes. Hydrogen bonds are shown by green dashed lines, the ity in hFXR. Sequence alignment indicated that most of electrostatic charge interaction is presented as a blue dashed line, the ligand contact residues identified in the crystallo- and the important cation-π interaction for FXR is depicted as a graphic structure of hFXR [60] are conserved between brown dashed line. The ligand is shown as ball and stick tnFXR and hFXR. However, one important hydrogen presentation and colored according to atom types (gray = carbon, bond interaction between the C-3 hydroxyl group of red-oxygen). CDCA and hFXR is not present for tnFXR because Tyr- 358 (hFXR) in helix 7 is substituted for by Phe in and less well by planar 5a bile alcohols, matching the tnFXR. endogenous primary bile salt profile of this species. There is an opening to solvent for the FXR LBP as Our experimental and molecular modelling studies of seen in crystallographic structures of human or rat tnFXR reveal a receptor that differs in multiple respects FXRs [59,60] and our homology models (Figure 10). from sea lamprey and zebrafish FXRs [9] but is instead The width of this opening is mostly controlled by two more similar to medaka FXRa [58] and mammalian residues - Met-262 and Ile-332 in human FXR. For FXRs. Homology modelling of tnFXR predicts an LBP tnFXR, the residue corresponding to hFXR Ile-332 is with a wide opening to solvent (wider than crystallo- alanine. Our model suggests that the LBP of tnFXR has graphic structure of hFXR [60]) and an overall architec- a wider opening to solvent than hFXR, which may ture that differs markedly from homology models of sea explain how tnFXR and not hFXR can accommodate lamprey and zebrafish FXRs, which are both predicted cyprinol 27-sulfate, a C27 bile acid that has a longer to have narrow LBPs that can accommodate planar bile side-chain and more hydroxyl groups than CDCA. The alcohols but cannot accommodate bent bile acids such pharmacophore model for tnFXR summarized in Figure as CDCA. Docking studies of lamprey and zebrafish 7C is consistent with the homology modelling and dock- FXRs predict good binding to planar 5a-bile alcohols ing results, in having multiple hydrophobic interactions, but not to bent 5b-bile acids, a prediction that is consis- a hydrogen bond acceptor, and negative ionizable fea- tent with reporter assay experiments [9]. For tnFXR, tures for the bound ligand. docking studies show that the FXR-selective agonist GW4064 and bile acid CDCA adopt orientations in the Discussion LBP similar to those observed in the crystallographic In this study, we characterize the FXR, VDR, and PXR of structures of GW4064 bound to hFXR [60] and rat FXR the green-spotted pufferfish Tetraodon nigriviridis,an bound to an analog of CDCA, respectively [59], lending actinopterygian fish that synthesizes mainly 5b-bile acids. further support to the conclusion that tnFXR has an This bile salt profile is similar to humans and most mam- overall LBP structure more similar to human and rat mals, as well as many other actinopterygian fish, some FXRs than to sea lamprey or zebrafish FXRs. reptiles, and some birds [13,23]. We find that the FXR Tetraodon FXR has similar ligand selectivity to medaka and PXR of Tetraodon (tnFXR and tnPXR) are activated FXRa (to which it shares high sequence identity in the by common bent-shaped 5b-bile acids such as CDCA LBD), including strong activation by GW4064 and the Krasowski et al. BMC Biochemistry 2011, 12:5 Page 14 of 19 http://www.biomedcentral.com/1471-2091/12/5

primary bile acid CDCA (both unconjugated and taurine- (Agnatha, which includes extant lampreys and hagfish) conjugated) [58]. Given that the activation of medaka and Cypriniformes (which includes zebrafish) [9,23]. FXRa upregulates the transcription of genes important Unlike lampreys and zebrafish, Tetraodon nigriviridis has in bile salt synthesis and transport, including the genes a bile salt profile of mainly bent 5b-bile acids such as for CYP7A1 and bile salt export protein as well as the CDCA, and would thus be predicted to have bile-salt- NHR repressor small heterodimer partner (SHP, NR0B2) regulating NHRs that can recognize Tetraodon endogen- [67], Tetraodon FXR seems likely to also be involved in ous bile acids and not the ‘ancestral’ bile alcohols that regulation of bile salt biology. Tetraodon FXR and comprise only a trace fraction of the Tetraodon bile salt medaka FXRa have a substantially different ligand selec- pool. These results lend further significant support to the tivity from sea lamprey FXR, which is not activated by hypothesis that structural changes to the LBD of FXR GW4064, CDCA (or other 5b-bile acids), or T-0901317 and PXR throughout evolution have paralleled cross- [9]. This is not surprising because the primary bile salts species changes in bile salt profile [6,7,10,11,66]. In of the sea lamprey (planar 5a-bile alcohols) are quite dif- Additional file 7 bile salt variation and NHR bile salt ferent from those of Tetraodon (CDCA, CA) and medaka sensitivity are overlaid on a fish phylogeny. So far, only a (CDCA, CA, and C27 bile acids) [23,57,58]. The FXRs of limited number of fish species have been characterized Tetraodon and medaka also differ significantly from an with respect to their FXRs, VDRs, and/or PXRs. As dis- FXR cloned and characterized from the little skate (Leu- cussed in Additional file 7 a number of fish groups would coraja erinacea, a cartilaginous fish), which was found to be of particular interest for future studies of NHRs, be insensitive to bile salts, even the 5b-bile alcohols (e.g., including the Elasmobranchii (sharks, skates, rays), Chi- 5b-scymnol sulfate) produced by the little skate and maeriformes (chimaerae), Myxiniformes (hagfish), and many other cartilaginous fish [68]. However, the skate lobe-finned fish (lungfish and coelacanths). FXR showed significant differences in sequence from Our present and previous studies [6,7,9,10] with VDRs other vertebrate FXRs, including novel insertions, and from various non-mammalian species have revealed there is the possibility that this receptor is actually ortho- some VDRs that are insensitive to bile acids (sea lam- logous to mammalian FXRb (NR1H5), which are acti- prey, African clawed frog) and others that are activated vated not by bile salts but instead by other steroidal by secondary bile acids such as LCA (human, mouse, compounds such as lanosterol [40]. chicken, medaka, Tetraodon). In humans and some Homology modelling of tnPXR predicts a receptor other mammals, enzymes from anaerobic bacteria in the with an LBP with a volume (1,230 Å3) significantly lar- caecum deconjugate and dehydroxylate primary bile ger than the estimated volumes of the LBPs of zebrafish acids such as CDCA, leading to secondary bile acids PXR (~1,000 Å3)andXenopus laevis PXRa (~860 Å3) such as LCA that can produce toxicity to the intestinal [9,66] but smaller than the LBP of hPXR in crystallo- and hepatobiliary tracts [15,21]. We and others have graphic structures [30,32-34]. The homology model and speculated that the ability of VDRs to bind secondary docking studies of tnPXR are consistent with our studies bile acids was acquired during vertebrate evolution as a of recombinant tnPXR in luciferase reporter assays that protective mechanism against the potential toxicity of show activation by a wider range of ligands than can poorly water-soluble secondary bile acids such as LCA activate either zebrafish or Xenopus PXRs. For example, [6,7,51,52]. One challenge to this hypothesis is that tnPXR is activated by a broad range of 5b-and5a-bile there has been little study to date of the secondary bile salts and steroids similar to hPXR. In contrast, studies salts (and the anaerobic bacterial intestinal flora that of recombinant zebrafish PXR show activation by a nar- could generate such bile salts) of non-mammalian spe- row range of planar 5a-bile alcohols [4,6,7]. The phar- cies. For example, it is not known whether Tetraodon or macophore model for tnPXR is similar to models of other actinopterygian fish species have physiologically hPXR [64,69,70] which have four hydrophobic features important amounts of secondary bile acids in the intest- along with a hydrogen bond acceptor, although the inal tract. With these caveats in mind, we summarize tnPXR pharmacophore had three hydrophobic features our studies of VDRs across different species. and a hydrogen bond donor, consistent with a some- The structure-activity and molecular modelling data what more restricted LBP limited to smaller ligands are both consistent with a tight, hydrophobic binding than hPXR [29,64]. Like hPXR, tnPXR has a pharmaco- pocket for bile acids in the human VDR LBP that can phore model markedly different from that of chicken bind bile acids with an oxo or single hydroxyl group at and zebrafish PXRs [29]. the C-3 position but not bile acids with substituents at Overall, the ligand selectivity of tnFXR and tnPXR con- the C-7 or C-12 positions (including unnatural bile trasts with that of sea lamprey FXR, zebrafish FXR, and acids with a single hydroxyl group on C-7 or C-12 and zebrafish PXR, which are activated better by 5a bile alco- no substituent on C-3). Molecular modeling predicts an hols, which are the main bile salts of jawless fish hVDR bile acid binding pocket that is predominantly Krasowski et al. BMC Biochemistry 2011, 12:5 Page 15 of 19 http://www.biomedcentral.com/1471-2091/12/5

hydrophobic but with polar features that permit hydro- Chemicals gen bond interaction with the 3a-hydroxy or 3-oxo The sources of chemicals were as follows: GW4064 group on the A ring of the bile acid and electrostatic (Sigma, St. Louis, MO, USA); T-0901317 (Axxora, San interactions with the bile acid side-chain as described Diego, CA, USA); 3a,7a,12a,24-tetrahydroxy-5a-cholan- above. Bile acids that have hydroxyl groups on steroid 24-sulfate (5a-petromyzonol sulfate; Toronto Research rings B and/or C (CDCA, DCA, CA, 7a-hydroxy-5b- Chemical, Inc., North York, ON, Canada); 1a,25-(OH)2- cholanic acid, 7b-hydroxy-5b-cholanic acid, and 12a- vitamin D3, Nuclear Receptor Ligand Library (76 com- hydroxy-5b-cholanic acid) do not interact favourably pounds known as ligands of various NHRs; BIOMOL with the more hydrophobic and sterically constrained International, Plymouth Meeting, PA, USA). All other portion of the hVDR bile acid binding pocket, consistent commercially purchased steroids and bile salts with the lack of activity of these bile acids in transacti- were obtained from Steraloids (Newport, RI, USA). vation assays. On the contrary, the unsubstituted and Norlithocholic acid, 3b-hydroxy-5a-cholan-24-oic acid, hydrophobic B, C, and D rings of LCA complement well 3b-hydroxy-5b-cholan-24-oic acid, 7a-hydroxy-5b-cho- the lipophilic portion of hVDR pocket and can form lan-24-oic acid, 7b-hydroxy-5b-cholan-24-oic acid, and numerous van der Waals interactions. We suggested 12a-hydroxy-5b-cholan-24-oic acid were generously that the unique physicochemical arrangement of the donated by the laboratory of A.F. Hofmann (University of hVDR ligand binding cavity provides the structural basis California - San Diego, La Jolla, CA, USA). for selective activation by LCA and its derivatives [9]. It is less clear, even with our mutagenesis studies, Cloning and expression of full-length VDRs exactly what changes mediate the cross-species differ- Full-length VDRa sequences from medaka (Oryzias ences in VDR pharmacology. Our studies rule out the latipes), zebrafish, and green spotted pufferfish were helix 1 - helix 3 ‘insertion’ domain as playing a role in identified within each species genome by conducting a the cross-species differences in activation by secondary generalized BLAST search using a query sequence bile acids. This insert is disordered in human and representing the P box domain of human VDR located mouse VDRs and shows highly variable sequences and within the highly conserved DNA binding region for the lengths across species [71-74]. Our site-directed muta- NHR superfamily. Full-length transcripts were subse- genesis experiments do confirm the previous prediction quently determined by identifying transcriptional start that Arg-274 and His-397 play a key role in interacting and stop codons for each gene. cDNAs containing a with bile acids [50,51]. Interestingly, Arg-274 and His- complete open reading frame for each gene were pro- 397 are conserved in the bile salt-insensitive xlVDR, duced by PCR using primer sets that spanned the entire indicating that other determinants underlie the differ- nucleic acid sequence for each gene including start and ences in bile salt pharmacology between xlVDR and stop codons. cDNAs were produced from extracts of bile-salt responsive VDRs. fish liver total RNA. Livers were homogenized with 1 mL RNA Bee (Tel-Test, Inc. Friendswood, TX, USA) Conclusions using a stainless steel Polytron homogenizer (Kinema- Our studies provide further important evidence of the tica, Lucerne, Switzerland) followed by cleanup and on- relationship between FXR, PXR, and VDR ligand selectiv- column DNase treatment using an RNeasy Mini Kit ity and cross-species variation in bile salt profiles. Zebra- (Qiagen, Valencia, CA, USA). fish and the green-spotted pufferfish Tetraodon nigriviridis RNA was eluted with 30 μl RNase-free water. RNA provide a clear contrast in having markedly different pri- quantity and quality were verified using an Agilent 2100 mary bile salt profiles (planar bile alcohols for zebrafish Bioanalyzer (Agilent Technologies, Santa Clara, CA, ® and sterically bent bile acids for the pufferfish) and corre- USA) and NanoDrop ND-1000 spectrophotometer spondingly contrasting receptor selectivity that matches (ThermoScientific, Wilmington, DE, USA). First strand the differences in endogenous ligands. In contrast, Tetrao- cDNA was made from total RNA (1-3 μg) and diluted don has a bile salt profile and receptor ligand selectivity with RNase-free water to a final volume of 10 μl, and 1 μl similar to humans. Our observations present an integrated oligo(dT)15 (500 μg/ml; Promega, Madison, WI, USA) picture of co-evolution of bile salt structure and the bind- and 1 μl 10 mM dNTPs were mixed with diluted RNA to ing pockets of three nuclear hormone receptors. yield a final volume of 20 μl. The mix was heated to 65°C for 5 min and chilled on ice for 2 min. Following centri- Methods fugation, 4 μl 5X first-strand buffer (Invitrogen, Carlsbad, Bile sample analysis CA, USA), 2 μl of 0.1 M DDT, and 1 μl RNase OUT Inhi- Bile salts from the green-spotted pufferfish were ana- bitor (40 U/μl; Invitrogen) were added to each reaction lyzed by HPLC and ESI/MS/MS using methods pre- and heated to 37°C. Following a 2 min incubation, 1 μl viously described [23,75]. Superscript Reverse Transcriptase (200 U/μl; Invitrogen) Krasowski et al. BMC Biochemistry 2011, 12:5 Page 16 of 19 http://www.biomedcentral.com/1471-2091/12/5

was added to each reaction and mRNA reverse tran- The LBD of chicken VDR was cloned from RNA scribed at 37°C for 1 h. All reverse transcription (RT) reac- extracted from the LMH cell line. The LBD of Xenopus tions were then inactivated by incubating at 70°C for 15 laevis VDR was cloned from RNA extracted from the min. cDNAs were stored at -20°C until PCR. PCR primers A6 cell line. The LBDs of tnFXR, tnVDR, and tnPXR for teleost VDRa were designed using PrimerQuest (Inte- were cloned from total RNA extracted from Tetraodon grated DNA Technologies, Coralville, IA, USA). PCR pri- liver using sequence information from the draft Tetrao- mers were flanked by restriction sites for incorporation don nigriviridis genome [76]. The LBDs of chVDR and transfer between appropriate cloning and expression (amino acid residues 113-451), xlVDR (amino acid resi- vectors. For each 25-μl PCR reaction, first-strand cDNAs dues 91-422), tnFXR (amino acid residues 180-463), were amplified using 2 μl (100-300 ng) first-strand cDNA, tnVDR (amino acid residues 91-426), and tnPXR (amino 9 μl RNase-free water, 0.75 μl10μM forward primer (0.3 acid residues 202-483) were inserted into the pM2- μM), 0.75 μl10μM reverse primer (0.3 μM), and 12.5 μl GAL4 vector to create GAL4/LBD chimeras. 2X Advantage Taq PCR Master Mix (Clontech, Mountain Site-directed mutations of mVDR were performed using View, CA, USA). PCR reaction conditions were: 95°C for the QuikChange II mutagenesis kit (Stratagene, La Jolla, 1.5 min followed by 35 cycles of 94°C for 15 s, 55°C for 30 CA, USA). All mutations were confirmed by sequencing of s, and 72°C for 1 min. PCR products for each teleost both DNA strands. To explore the importance of the H1- VDRa sequence were cloned into the TA cloning vector H3 insert in ligand selectivity, four constructs were created. pCR2.1 (Invitrogen, Carlsbad, CA) as per manufacturer’s Thus, we generated hVDR/Δins that lacked amino acid suggestions. VDR’s were subsequently excised from pCR residues 165-218 (hVDR/Δins) and a hVDR construct 2.1 using EcoRI and BamHI and inserted unidirectionally where the insert domain was replaced with the insertion into the expression vector pSG5. Proper orientation of domain from bile acid-insensitive xlVDR (hVDR/xlVDRins; VDRa’s within the vector was confirmed by PCR screen- hVDR residues 158-223 replaced by xlVDR residues 160- ing and sequencing in both directions. 218). The analogous constructs were also created for HepG2 cells were cultured in T75 flasks with vented xlVDR: one lacking the insertion domain (xlVDR/Δins; caps (Corning, Corning, NY) using MEM containing missing amino acid residues 166-211) and the other head-inactivated fetal bovine serum (10%), 1X sodium (xlVDR/hVDRins; xlVDR residues 160-218 replaced by pyruvate, 1X nonessential amino acids, and 1% penicil- hVDR residues 158-223). The chimeras were generated by lin/streptomycin. Cells were maintained at 37°C with 5% synthesis of double-stranded DNA (Genscript, Piscataway, CO2 and split regularly at 70-80% confluency. For tran- NJ) which was inserted into the pM2-GAL4 vector. sient transfection, HepG2 cells were seeded at a density of 1 × 105 cells per well in 24 well plates in antibiotic- Cell lines free MEM. Wells were transfected over night with either The creation of a HepG2 (human liver) cell line stably empty pSG5 vector (control), medaka VDRa,zebrafish expressing the human Na+-taurocholate cotransporter VDRa,orTetraodon VDRa. Each well contained 299 ng (NTCP; SLC10A1) has been previously reported [7]. pSG5-VDRa foreither(medaka,zebrafish,orTetrao- HepG2-NTCP cells were grown in modified Eagle’s med- don), 64 ng hCYP3A4-Luc reporter construct (XREM), ium-a containing 10% fetal bovine serum and 1% penicil- and 15 ng of the Renilla normalizing plasmid (pRL- lin/streptomycin. The cells were grown at 37°C in 5% CMV). Luciferase reporter assay experiments were CO2. The chicken LMH hepatoma cell line (ATCC, Man- performed as previously described [58]. assus, VA, USA) was grown in Waymouth’s MB 752/1 Plasmids containing human organic anion transporting medium (ATCC) with 10 fetal bovine serum. The Xeno- polypeptide (Oatp1a1; Slco1a1), as well as the reporter pus laevis A6 kidney cell line (ATCC) was grown in 75% construct tk-UAS-Luc and the ‘empty’ vector PM2, were NCTC 109 medium, 15% distilled water, and 10% fetal generously provided by SA Kliewer, JT Moore, and LB bovine serum at 26°C in 2% CO2. Except as noted above, Moore (GlaxoSmithKline, Research Triangle Park, NC, all media and media supplements for the HepG2 and A6 USA). To permit comparison between species and to cell lines were obtained from Invitrogen (Carlsbad, CA, avoid mismatching of non-mammalian receptors with USA). Co-transfections and transactivations assays were mammalian retinoid X receptor, co-factors, and chroma- performed as previously described [77]. The maximal tin remodeling factors, all receptors were studied as activators and their concentrations were as follows: LBD/GAL4 chimeras. For the GAL4/LBD expression chVDR, xlVDR, and tnVDR - 200 nM 1a,25-(OH)2- constructs, the reporter plasmid is tk-UAS-Luc, which vitamin D3;tnFXR-5μM GW4064; tnPXR - 50 μM contains GAL4 DNA binding elements driving luciferase 5a-androstan-3a-ol. All comparisons to maximal activa- expression. The cloning of LBDs from hFXR, hVDR, tors were done within the same microplate. Luciferase mouse VDR (mVDR), and zebrafish VDR has been pre- data were normalized to the internal b-galactosidase con- viously reported [6,7,9]. trol and represent means ± SD of the assays. Krasowski et al. BMC Biochemistry 2011, 12:5 Page 17 of 19 http://www.biomedcentral.com/1471-2091/12/5

Common features pharmacophore development and Additional file 7: Fish phylogeny, bile salt variation, and nuclear database screening hormone receptor ligand specificity. Computational molecular modeling studies were carried out using Catalyst™in Discovery Studio 2.5.5 (Accelrys, San Diego, CA, USA) by methods previously described Acknowledgements [29]. Common features pharmacophores were developed MDK is supported by K08-GM074238 from the National Institutes of Health. based on the compounds tested against tnFXR, tnPXR SWK is supported by the National Science Foundation NSF0842510. SE kindly and tnVDR using the HipHop method [78,79]. A HipHop acknowledges Accelrys for providing Discovery Studio for use in this work. pharmacophore attempts to describe the arrangement of Author details key features that are important for biological activity. 1Department of Pathology, University of Iowa Hospitals and Clinics, Iowa City, IA, 52242, USA. 2Department of Pathology, University of Pittsburgh, Pittsburgh, PA, 15261 USA. 3Department of Pharmacology, Robert Wood Molecular modeling studies Johnson Medical School, University of Medicine & Dentistry of New Jersey, A structural model of the LBDs of the tnFXR and Piscataway, NJ 08854, USA. 4Department of Medicine, University of California 5 tnPXR were constructed using the Molecular Operat- - San Diego, La Jolla, CA, 92093 USA. Department of Environmental and Molecular Toxicology, North Carolina State University, Raleigh, NC 27695, ing Environment (MOE; Chemical Computating USA. 6Collaborations in Chemistry, Jenkintown, PA 19046, USA. 7Department Group, Montreal, Canada) Homology Model module. of Pharmaceutical Sciences, University of Maryland, Baltimore, MD 21201, The modeling template used for FXR is the published USA. crystal structure of rat FXR in complex with 6a-ethyl- Authors’ contributions chenodeoxycholic acid (PDB ID = 1OSV) [59] (this NA and SE performed molecular modelling studies, including those that structure was chosen because it has bound bile acid provided the basis for the site-directed mutagenesis experiments. LRH performed the analysis of bile salts of the green-spotted pufferfish. EMK and unlike available crystal structures of hFXR). The tem- SWK performed the cloning, expression, and analysis of full-length VDRs in plate for PXR is hPXR co-crystallized with SR12813 three actinopterygian fish species. EJR and MDK cloned Tetraodon FXR, (PDB ID = 1nrl) [33]. Several energy minimization- Tetraodon PXR, chicken VDR, Xenopus laevis VDR, and Tetraodon VDR, as well as all site-directed mutants of mouse VDR. MDK performed the function based refinement procedures were implemented on the assays of all GAL4/LBD chimeric receptors and drafted the manuscript. All initial model, and the quality of the final model was authors contributed to, read, and approved the final manuscript. confirmed by the WHATIF-Check program. Estimation Received: 20 September 2010 Accepted: 3 February 2011 of the volume of the LBP for the crystallographic Published: 3 February 2011 structures and homology models described above were determined using CASTp http://sts.bioengr.uic.edu/ References castp/calculation.php [80]. 1. Huang P, Chandra V, Rastinejad F: Structural overview of the nuclear receptor superfamily: insights into physiology and therapeutics. Annu Molecular docking studies were performed by the Rev Physiol 2010, 72:247-272. GOLD docking program [81]. In addition to the docking 2. McEwan IJ: Nuclear receptors: one big family. Methods Mol Biol 2009, studies with tnPXR and tnFXR, LCA was docked into 505:3-18. 3. Steinmetz ACU, Renaud J-P, Moras D: Binding of ligands and activation of the crystal structure for human VDR (PDB accession transcription by nuclear receptors. Annu Rev Biophys Biomol Struct 2001, number 1DB1). During the docking process, the protein 30:329-359. was held fixed while full conformational flexibility was 4. Moore LB, Maglich JM, McKee DD, Wisely B, Willson TM, Kliewer SA, Lambert MH, Moore JT: Pregnane X receptor (PXR), constitutiveandrostanereceptor(CAR), allowed for ligands. For each ligand, 30 independent and benzoate X receptor (BXR) define three pharmacologically distinct docking runs were performed to achieve the consensus classes of nuclear receptors. Mol Endocrinol 2002, 16:977-986. orientation in the LBP. 5. Moore LB, Parks DJ, Jones SA, Bledsoe RK, Consler TG, Stimmel JB, Goodwin B, Liddle C, Blanchard SG, Willson TM, et al: Orphan nuclear receptors constitutive androstane receptor and pregnane X receptor Additional material share xenobiotic and steroid ligands. J Biol Chem 2000, 275:15122-15127. 6. Krasowski MD, Yasuda K, Hagey LR, Schuetz EG: Evolutionary selection across the nuclear hormone receptor superfamily with a focus on the Additional file 1: Analyses of biliary bile salts from the green- NR1I subfamily (vitamin D, pregnane X, and constitutive androstane spotted pufferfish. receptors). Nucl Recept 2005, 3:2. Additional file 2: Activation data for Tetraodon FXR and PXR by bile 7. Krasowski MD, Yasuda K, Hagey LR, Schuetz EG: Evolution of the pregnane salts, steroids, and synthetic ligands. X receptor: adaptation to cross-species differences in biliary bile salts. Mol Endocrinol 2005, 19:1720-1739. Additional file 3: Activation data for VDRs, including the site- 8. Iyer M, Reschly EJ, Krasowski MD: Functional evolution of the pregnane X directed mutants of mouse VDR. receptor. Expert Opin Drug Metab Toxicol 2006, 2:381-387. Additional file 4: The ligand binding pocket of human VDR. Surface 9. Reschly EJ, Ai N, Ekins S, Welsh WJ, Hagey LR, Hofmann AF, Krasowski MD: representation of human VDR ligand binding pocket. Evolution of the bile salt nuclear receptor FXR in vertebrates. J Lipid Res Additional file 5: Amino acid sequence alignment for VDRs 2008, 49:1577-1587. (including the four chimeric constructs involving the H1-H3 10. Reschly EJ, Bainy ACD, Mattos JJ, Hagey LR, Bahary N, Mada SR, Ou J, insertion domain) and human PXR. Venkataramanan R, Krasowski MD: Functional evolution of the vitamin D and pregnane X receptors. BMC Evol Biol 2007, 7:222. Additional file 6: Data for pharmacophore modeling of Tetraodon 11. Reschly EJ, Krasowski MD: Evolution and function of the NR1I nuclear nuclear hormone receptors. hormone receptor subfamily (VDR, PXR, and CAR) with respect to Krasowski et al. BMC Biochemistry 2011, 12:5 Page 18 of 19 http://www.biomedcentral.com/1471-2091/12/5

metabolism of xenobiotics and endogenous compounds. Curr Drug 37. Makishima M, Okamato AY, Repa JJ, Tu H, Learned RM, Luk A, Hull MV, Metab 2006, 7:349-365. Lustig KD, Mangelsdorf DJ, Shan B: Identification of a nuclear receptor for 12. Schuetz EG, Strom S, Yasuda K, Lecureur V, Assem M, Brimer C, Lamba J, bile acids. Science 1999, 284:1362-1365. Kim RB, Ramachandran V, Komoroski BJ, et al: Disrupted bile acid 38. Parks DJ, Blanchard SG, Bledsoe RK, Chandra G, Consler TG, Kliewer SA, homeostasis reveals an unexpected interaction among nuclear hormone Stimmel JB, Willson TM, Zavacki AM, Moore DD, et al: Bile acids: natural receptors, transporters, and cytochrome P450. J Biol Chem 2001, ligands for an orphan nuclear receptor. Science 1999, 284:1365-1368. 276:39411-39418. 39. Wang H, Chen J, Hollister K, Sowers LC, Forman BM: Endogenous bile 13. Hofmann AF, Hagey LR, Krasowski MD: Bile salts of vertebrates: structural acids are ligands for the nuclear receptor FXR/BAR. Mol Cell 1999, variation and possible evolutionary significance. J Lipid Res 2010, 3:543-553. 51:226-246. 40. Otte K, Kranz H, Kober I, Thompson P, Hoefer M, Haubold B, Remmel B, 14. Hofmann AF, Eckmann L: How bile acids confer gut mucosal protection Voss H, Kaiser C, Albers M, et al: Identification of farnesoid X receptor b as against bacteria. Proc Natl Acad Sci USA 2006, 103:4333-4334. a novel mammalian nuclear receptor sensing lanosterol. Mol Cell Biol 15. Hofmann AF, Hagey LR: Bile acids: chemistry, pathochemistry, biology, 2003, 23:864-872. pathobiology, and therapeutics. Cell Mol Life Sci 2008, 65:2461-2483. 41. Seo Y-W, Sanyal S, Kim H-J, Won DH, An J-Y, Amano T, Zavacki AM, 16. Haslewood GAD: Bile salt evolution. J Lipid Res 1967, 8:535-550. Kwon H-B, Shi Y-B, Kim W-S, et al: FOR, a novel orphan nuclear receptor 17. Moschetta A, Xu F, Hagey LR, van Berge Henegouwen GP, van Erpecum KJ, related to farnesoid X receptor. J Biol Chem 2002, 277:17836-17844. Brouwers JF, Cohen JC, Bierman M, Hobbs HH, Steinbach JH, et al: A 42. Dusso AS, Brown AJ, Slatopolsky E: Vitamin D. Am J Physiol Renal Physiol phylogenetic survey of biliary lipids in vertebrates. J Lipid Res 2005, 2005, 289:F8-F28. 46:2221-2232. 43. Holick MF: Evolution and function of vitamin D. Recent Results Cancer Res 18. Une M, Hoshita T: Natural occurrence and chemical synthesis of bile 2003, 164:3-28. alcohols, higher bile acids, and short side chain bile acids. Hiroshima J 44. Kira M, Kobayashi T, Yoshikawa K: Vitamin D and the skin. J Dermatol 2003, Med Sci 1994, 43:37-67. 30:429-437. 19. Norlin M, Wikvall K: Enzymes in the conversion of cholesterol into bile 45. Moro JR, Iwata M, von Andriano UH: Vitamin effects on the immune acids. Curr Mol Med 2007, 7:199-218. system: vitamins A and D take centre stage. Nat Rev Immunol 2008, 20. Russell DW: The enzymes, regulation, and genetics of bile acid synthesis. 8:685-698. Annu Rev Biochem 2003, 72:137-174. 46. Bertrand S, Thisse B, Tavares R, Sachs L, Chaumot A, Bardet PL, Escriva H, 21. Hofmann AF: Detoxification of lithocholic acid, a toxic bile acid: Duffraisse M, Marchand O, Safi R, et al: Unexpected novel relational links relevance to drug hepatotoxicity. Drug Metab Rev 2004, 36:703-722. uncovered by extensive developmental profiling of nuclear receptor 22. Goto T, Holzinger F, Hagey LR, Cerrè C, Ton-Nu H-T, Schteingart CD, expression. PLoS Genet 2007, 3:e188. Steinbach JH, Shneider BL, Hofmann AF: Physicochemical and 47. Jaillon O, Aury J-M, Brunet F, Petit J-L, Strange-Thomann N, Mauceli E, physiological properties of 5a-cyprinol sulfate, the toxic bile salt of Bouneau L, Fischer C, Ozouf-Costaz C, Bernot A, et al: Genome duplication cyprinid fish. J Lipid Res 2003, 44:1643-1651. in the telost fish Tetraodon nigroviridis reveals the early vertebrate 23. Hagey LR, Møller PR, Hofmann AF, Krasowski MD: Diversity of bile salts in proto-karyotype. Nature 2004, 431:946-956. fish and amphibians: evolution of a complex biochemical pathway. 48. Maglich JM, Caravella JA, Lambert MH, Willson TM, Moore JT, Physiol Biochem Zool 2010, 83:308-321. Ramamurthy L: The first completed genome sequence from a teleost fish 24. Hagey LR, Vidal N, Hofmann AF, Krasowski MD: Evolutionary diversity of (Fugu rubripes) adds significant diversity to the nuclear receptor bile salts in reptiles and mammals, including analysis of ancient human superfamily. Nucleic Acids Res 2003, 31:4051-4058. and extinct giant ground sloth coprolites. BMC Evol Biol 2010, 10:133. 49. Howarth DL, Law SH, Barnes B, Hall JM, Hinton DE, Moore L, Maglich JM, 25. Hagey LR, Vidal N, Hofmann AF, Krasowski MD: Complex evolution of bile Moore JT, Kullman SW: Paralogous vitamin D receptors in teleosts: salts in birds. Auk 2010, 127:820-831. transition of nuclear receptor function. Endocrinology 2008, 149:2411-2422. 26. Kliewer SA, Willson TM: Regulation of xenobiotic and bile acid metabolism 50. Adachi R, Honma Y, Masuno H, Karana K, Shimomura I, Yamada S, by the nuclear pregnane X receptor. J Lipid Res 2002, 43:359-364. Makishima M: Selective activation of vitamin D receptor by lithocholic 27. Orans J, Teotico DG, Redinbo MR: The nuclear xenobiotic receptor PXR: acid acetate, a bile acid derivative. J Lipid Res 2005, 46:46-57. recent insights and new challenges. Mol Endocrinol 2005, 19:2891-2900. 51. Adachi R, Shulman AI, Yamamoto K, Shimomura I, Yamada S, 28. Sonoda J, Xie W, Rosenfeld JM, Barwick JL, Guzelian PS, Evans RM: Mangelsdorf DJ, Makishima M: Structural determinants for vitamin D Regulation of a xenobiotics sulfonation cascade by nuclear pregnane X responses to endocrine and xenobiotic signals. Mol Endocrinol 2004, receptor (PXR). Proc Natl Acad Sci USA 2002, 99:13801-13806. 18:43-52. 29. Ekins S, Reschly EJ, Hagey LR, Krasowski MD: Evolution of pharmacologic 52. Makishima M, Lu TT, Xie W, Whitfield GK, Domoto H, Evans RM, specificity in the pregnane X receptor. BMC Evol Biol 2008, 8:103. Haussler MR, Mangelsdorf DJ: Vitamin D receptor as an intestinal bile acid 30. Chrencik JE, Orans J, Moore LB, Xue Y, Peng L, Collins JL, Wisely GB, sensor. Science 2002, 296:1313-1316. Lambert MH, Kliewer SA, Redinbo MR: Structural disorder in the complex 53. Chow EC, Maeng HJ, Liu S, Khan AA, Groothuis GM, Pang KS: 1alpha,25- of human PXR and the macrolide antibiotic rifampicin. Mol Endocrinol dihydroxyvitamin D(3) triggered vitamin D receptor and farnesoid X 2005, 19:1125-1134. receptor-like effects in rat intestine and liver in vivo. Biopharm Drug 31. Watkins RE, Davis-Searles PR, Lambert MH, Redinbo MR: Coactivator Dispos 2009. binding promotes the specific interaction between ligand and the 54. Matsubara T, Yoshinari K, Aoyama K, Sugawara M, Sekiya Y, Nagata K, pregnane X receptor. J Mol Biol 2003, 331:815-828. Yamazoe Y: Role of vitamin D receptor in the lithocholic acid-mediated 32. Watkins RE, Maglich JM, Moore LB, Wisely GB, Noble SM, Davis-Searles PR, CYP3A induction in vitro and in vivo. Drug Metab Dispos 2008, Lambert MH, Kliewer SA, Redinbo MR: 2.1 Å crystal structure of human 36:2058-2063. PXR in complex with the St. John’s wort compound hyperforin. 55. Pavek P, Pospechova K, Svecova L, Syrova Z, Stejskalova L, Blazkova J, Dvorak Z, Biochemistry 2003, 42:1430-1438. Blahos J: Intestinal cell-specific vitamin D receptor (VDR)-mediated 33. Watkins RE, Wisely GB, Moore LB, Collins JL, Lambert MH, Williams SP, Willson TM, transcriptional regulation of CYP3A4 gene. Biochem Pharmacol 2009. Kliewer SA, Redinbo MR: The human nuclear xenobiotic receptor PXR: structural 56. Thummel KE, Brimer C, Yasuda K, Thottassery J, Senn T, Lin Y, Ishizuka H, determinants of directed promiscuity. Science 2001, 292:2329-2333. Kharasch E, Schuetz J, Schuetz E: Transcriptional control of intestinal 34. Xue Y, Chao E, Zuercher WJ, Willson TM, Collins JL, Redinbo MR: Crystal cytochrome P-4503A by 1a,25-dihydroxy vitamin D3. Mol Pharmacol 2001, structure of PXR-T1317 complex provides a scaffold to examine the 60:1399-1406. potential for receptor antagonism. Bioorg Med Chem 2007, 15:2156-2166. 57. Hagey LR, Iida T, Ogawa S, Kakiyama G, Asahina K, Mushiake K, Goto T, 35. Xue Y, Moore LB, Orans J, Peng L, Bencharit S, Kliewer SA, Redinbo MR: Mano N, Goto J, Krasowski MD, et al: Major biliary bile acids of the Crystal structure of PXR-estradiol complex provides insights into Medaka (Oryzias latipes): 25R- and 25S epimers of 3α,7α,12α-trihydroxy- endobiotic recognition. Mol Endocrinol 2007, 21:1028-1038. 5β-cholestanoic acid. Zool Sci 2010, 27:565-573. 36. Kalaany NY, Mangelsdorf DJ: LXRs and FXR: the Yin and Yang of 58. Howarth DL, Hagey LR, Law SHW, Ai N, Krasowski MD, Ekins S, Moore JT, cholesterol and fat metabolism. Annu Rev Physiol 2006, 68:159-191. Kollitz EM, Hinton DE, Kullman SW: Two farnesoid X receptor a isoforms in Krasowski et al. BMC Biochemistry 2011, 12:5 Page 19 of 19 http://www.biomedcentral.com/1471-2091/12/5

Japanese medaka (Orzias latipes) are differentially activated in vitro. 81. Jones G, Willett P, Glen RC, Leach AR, Taylor R: Development and Aquat Toxicol 2010, 98:245-255. validation of a genetic algorithm for flexible docking. J Mol Biol 1997, 59. Mi L-Z, Devarakonda S, Harp JM, Han Q, Pellicciari R, Willson TM, 267:727-748. Khorasanizadeh S, Rastinejad F: Structural basis for bile acid binding and activation of the nuclear receptor FXR. Mol Cell 2003, 11:1093-1100. doi:10.1186/1471-2091-12-5 60. Downes M, Verdecia MA, Roecker AJ, Hughes R, Hogenesch JB, Kast- Cite this article as: Krasowski et al.: The evolution of farnesoid X, vitamin Woelbern HR, Bowman ME, Ferrer J-L, Anisfeld AM, Edwards PA, et al: A D, and pregnane X receptors: insights from the green-spotted pufferfish chemical, genetic, and structural analysis of the nuclear bile acid (Tetraodon nigriviridis) and other non-mammalian species. BMC receptor FXR. Mol Cell 2003, 11:1079-1092. Biochemistry 2011 12:5. 61. Houck KA, Borchert KM, Hepler CD, Thomas JS, Bramlett KS, Michael LF, Burris TP: T0901317 is a dual LXR/FXR agonist. Mol Genet Metab 2004, 83:184-187. 62. Mitro N, Vargas L, Romeo R, Koder A, Saez E: T0901317 is a potent PXR ligand: implications for the biology ascribed to LXR. FEBS Lett 2007, 581:1721-1726. 63. Blumberg B, Kang H, Bolado J, Chen H, Craig AG, Moreno TA, Umesano K, Perlmann T, De Robertis EM, Evans RM: BXR, an embryonic orphan nuclear receptor activated by a novel class of endogenous benzoate metabolites. Genes Dev 1998, 12:1269-1277. 64. Ekins S, Erickson JA: A pharmacophore for human pregnane X receptor ligands. Drug Metab Dispos 2002, 30:96-99. 65. Ekins S, Kortagere S, Iyer M, Reschly EJ, Lill MA, Redinbo MR, Krasowski MD: Challenges predicting ligand-receptor interactions of promiscuous proteins: the nuclear receptor PXR. PLoS Comput Biol 2009, 5:e1000594. 66. Krasowski MD, Ni A, Hagey LR, Ekins S: Evolution of promiscuous nuclear hormone receptors: LXR, FXR, VDR, PXR, and CAR. Mol Cell Endocrinol 2010. 67. Howarth DL, Law SH, Law JM, Mondon JA, Kullman SW, Hinton DE: Exposure to the synthetic FXR agonist GW4064 causes alterations in gene expression and sublethal hepatotoxicity in eleutheroembryo medaka (Oryzias latipes). Toxicol Appl Pharmacol 2010, 243:111-121. 68. Cai SY, Xiong L, Wray CG, Ballatori N, Boyer JL: The farnesoid X receptor, FXRa/NR1H4, acquired ligand specificity for bile salts late in vertebrate evolution. Am J Physiol Regul Integr Comp Physiol 2007, 293:R1400-R1409. 69. Bachmann K, Patel H, Batayneh Z, Slama J, White D, Posey J, Ekins S, Gold D, Sambucetti L: PXR and the regulation of apoA1 and HDL- cholesterol in rodents. Pharmacol Res 2004, 50:237-246. 70. Schuster D, Langer T: The identification of ligand features essential for PXR activation. J Med Chem 2005, 45:431-439. 71. Ciesielski F, Rochel N, Mitschler A, Kouzmenko A, Moras D: Structural investigation of the ligand binding domain of the zebrafish VDR in complexes with 1alpha,25(OH)2D3 and Gemini: purification, crystallization and preliminary X-ray diffraction analysis. J Steroid Biochem Mol Biol 2004, 89-90:55-59. 72. Ciesielski F, Rochel N, Moras D: Adaptability of the Vitamin D nuclear receptor to the synthetic ligand Gemini: remodelling the LBP with one side chain rotation. J Steroid Biochem Mol Biol 2007, 103:235-242. 73. Rochel N, Tocchini-Valentini G, Egea PF, Juntunen K, Garnier J-M, Vihko P, Moras D: Functional and structural characterization of the insertion region in the ligand binding domain of the vitamin D nuclear receptor. Eur J Biochem 2001, 268:971-979. 74. Rochel N, Wurtz JM, Mitschler A, Klaholz B, Moras D: The crystal structure of the nuclear receptor for vitamin D bound to its natural ligand. Mol Cell 2000, 5:173-179. 75. Rossi SS, Converse JL, Hofmann AF: High pressure liquid chromatography analysis of conjugated bile acids in human bile: simultaneous resolution of sulfated and unsulfated lithocholyl amidates and the common conjugated bile acids. J Lipid Res 1987, 28:589-595. 76. Ensembl genome browser. [http://www.ensembl.org]. Submit your next manuscript to BioMed Central 77. Reschly EJ, Ai N, Welsh WJ, Ekins S, Hagey LR, Krasowski MD: Ligand and take full advantage of: specificity and evolution of liver X receptors. Journal of Steroid Biochemistry and Molecular Biology 2008, 110:83-94. 78. Clement OO, Mehl AT: HipHop: pharmacophore based on multiple • Convenient online submission common-feature alignments. In Pharmacophore perception, development, • Thorough peer review and use in drug design. Edited by: Guner OF. San Diego: IUL; 2000:69-84. • No space constraints or color figure charges 79. Diao L, Ekins S, Polli JE: Novel inhibitors of human organic cation/ carnitine transporter (hOCTN2) via computational modeling and in vitro • Immediate publication on acceptance testing. Pharm Res 2009, 26:1890-1900. • Inclusion in PubMed, CAS, Scopus and Google Scholar 80. Liang J, Edelsbrunner H, Woodward C: Anatomy of protein pockets and • Research which is freely available for redistribution cavities: measurement of binding site geometry and implications for ligand design. Protein Sci 1998, 7:1884-1897. Submit your manuscript at www.biomedcentral.com/submit